The evidence scoreboard

HRV accuracy, device by device

Five wearables worn simultaneously against a single-lead ECG, every night, for 536 nights. This is the cleanest comparison in the whole category.

By Stephen V.Published August 11, 2026

#ad

We earn a commission when you buy through our Amazon links, at no extra cost to you. It never decides a ranking — the rubric that does is published in full, so you can check us rather than take our word for it. How this works.

The short answer

Against a single-lead ECG reference across 536 nights, the Oura Ring 4 recorded the lowest HRV error (5.96% MAPE, concordance 0.99), followed by the Oura Ring 3 (7.15%), Whoop 4.0 (8.17%), Garmin fēnix 6 (10.52%) and Polar Grit X Pro (16.32%). The reference device was a Polar H10 chest strap.

Why this study is unusually good

Cross-device accuracy comparisons are usually worthless, because the devices are tested on different people, on different nights, in different conditions. Dial et al. (2025) avoided all of that by having 13 participants wear all five wearables at the same time, every night, against the same ECG reference — 536 nights in total.

Same body, same night, same sleep, same reference. When the numbers differ, the device is the reason. That is a rare and valuable design in this category, and it is why we treat this as the strongest HRV evidence available.

What the numbers mean in practice

MAPE is the average percentage by which the device missed the reference. Concordance correlation combines agreement and precision into a single figure where 1.0 is perfect.

A 6% error on HRV is genuinely good — night-to-night HRV swings by far more than that in normal life, so the signal survives the noise. A 16% error with a ±24% standard deviation is a different situation: the measurement error is now comparable to the physiological change you are trying to detect, which makes a trend line hard to interpret.

That spread is the part to read carefully. Two devices with similar averages can be very different instruments if one is consistent and the other is not.

The Garmin exclusion is worth a paragraph

Garmin's resting heart rate could not be compared at all. The authors write that the device reports "the lowest 30 min average in a 24 h period" but does not timestamp which 30 minutes, so there was no way to align it with the reference data.

That is not an accuracy finding — it is a transparency finding, and it is arguably more interesting. A number you cannot audit is not a number you can check, and the whole premise of this site is that checkability is what separates a ranking from a vibe.

The limits, in the authors' own terms

Thirteen participants. Many nights, few people, all apparently healthy adults. The authors explicitly flag generalizability to sleep and cardiovascular disorders as untested.
Unequal night counts.The Oura Gen 4 arm rests on 139 nights against the Gen 3's 470. Fewer nights, wider uncertainty.
Nocturnal only. Nothing here describes HRV measured during the day or around training.
Moving targets. The authors note these algorithms are updated periodically and recommend continued re-validation — which means every figure above has a shelf life.

What this changes about buying

It is the reason our HRV rankingputs a chest strap first and a ring second, and the reason we do not rank a Polar watch as an HRV device anywhere on this site — the one Polar watch that has been measured for nocturnal HRV came last of five, while the same company's chest strap is the instrument everything else was measured against.

It is also why we say nothing confident about readiness scores. The input has now been measured for five devices. The proprietary model sitting on top of it has been measured for none of them.

The record

What has actually been measured

Every row cites a numbered source below, or states outright that no validation study exists. There is no third state.
Device	Metric	What was actually measured	Evidence
Oura Gen 4	HRV (RMSSD), nocturnal	Concordance 0.99, MAPE 5.96 ± 5.12% across 139 nights — best of five devices [1]	Peer-reviewed validation
Oura Gen 3	HRV (RMSSD), nocturnal	Concordance 0.97, MAPE 7.15 ± 5.48% across 470 nights [1]	Peer-reviewed validation
Whoop 4.0	HRV (RMSSD), nocturnal	Concordance 0.94, MAPE 8.17 ± 10.49% across 289 nights [1]	Peer-reviewed validation
Garmin fēnix 6	HRV (RMSSD), nocturnal	Concordance 0.87, MAPE 10.52 ± 8.63% across 150 nights — described by the authors as poor agreement [1]	Peer-reviewed validation
Polar Grit X Pro	HRV (RMSSD), nocturnal	Concordance 0.82, MAPE 16.32 ± 24.39% across 206 nights — weakest of the five devices [1]	Peer-reviewed validation
Oura Gen 4	Resting heart rate, nocturnal	Concordance 0.98, MAPE 1.94 ± 2.51% [1]	Peer-reviewed validation
Oura Gen 3	Resting heart rate, nocturnal	Concordance 0.97, MAPE 1.67 ± 1.54% — lowest error of any device measured [1]	Peer-reviewed validation
Whoop 4.0	Resting heart rate, nocturnal	Concordance 0.91, MAPE 3.00 ± 2.15% [1]	Peer-reviewed validation
Polar Grit X Pro	Resting heart rate, nocturnal	Concordance 0.86, MAPE 2.71 ± 2.75% [1]	Peer-reviewed validation
Garmin fēnix 6	Resting heart rate, nocturnal	Not comparable. The authors excluded it because the 30-minute window Garmin uses to derive resting heart rate is not timestamped, so it could not be aligned with the reference.	No validation study exists
Polar H10	HRV, reference	Used as the criterion measure — a single-lead ECG sampled at 1000 Hz. Its own error is therefore unmeasured here. [1]	Peer-reviewed validation
Every readiness or recovery score	Score validity	Unmeasured. The authors state there is little transparency into what metrics feed each device's own readiness or recovery score, and that those algorithms are periodically updated.	No validation study exists
Oura Ring 4 (current retail model)	HRV, this generation	The study tested Oura Gen 3 and Gen 4 hardware; we do not assume any result transfers to a subsequent generation, here or anywhere else on this site.	No validation study exists
RingConn, Ultrahuman, Fitbit, Apple Watch	HRV, nocturnal	Not tested. No comparable HRV validation of these devices that we could find.	No validation study exists

Questions

Frequently asked

Which wearable has the most accurate HRV?

Of the five measured in the 2025 ECG validation, the Oura Ring — 5.96% MAPE and 0.99 concordance for the Gen 4, 7.15% for the Gen 3. Whoop 4.0 followed at 8.17%, then Garmin fēnix 6 at 10.52% and Polar Grit X Pro at 16.32%. The reference itself was a Polar H10 chest strap.

Is Garmin HRV accurate?

The one measurement we found puts a fēnix 6 at 10.52 ± 8.63% MAPE with a concordance of 0.87 — described by the study authors as poor agreement, and behind Oura and WHOOP. Separately, Garmin's resting heart rate could not be compared at all because the 30-minute window it uses is not timestamped.

Why is a chest strap used as the reference?

Because it measures the heart's electrical activity directly and can time individual beats to the millisecond, which is exactly what HRV arithmetic requires. Optical sensors infer the pulse from light scattered through tissue, which adds error at every step. Researchers use the electrical measurement as the yardstick for that reason.

Does a good HRV number mean the recovery score is good?

No, and the study authors say so directly: there is little transparency into what feeds each device's readiness or recovery score, and those algorithms change over time. A device can measure HRV accurately and still produce a score that has never been validated by anyone.

Show your working

Sources

Every one of these was read in full before it was cited. Where we could not access a paper, we do not quote its numbers.

[1]Peer-reviewed validation2025
Dial, Hollander, Vatne, Emerson, Edwards & Hagen, “Validation of nocturnal resting heart rate and heart rate variability in consumer wearables”, Physiological Reports 13(16):e70527
DOI 10.14814/phy2.70527. n = 13 healthy adults (7 male, 6 female, 33.2 ± 8.6 years) across 536 nights, each wearing an Oura Gen 3, Oura Gen 4, Polar Grit X Pro, Garmin fēnix 6 and Whoop 4.0 simultaneously against a Polar H10 single-lead ECG reference sampled at 1000 Hz. Authors’ stated limits: healthy adults only, no atrial-fibrillation population, proprietary and periodically-updated algorithms, and unequal night counts per device. The Garmin was dropped from the resting-heart-rate comparison because the 30-minute window it uses is not timestamped.
[2]Manufacturer specification2026
Polar — H10 heart rate sensor published specifications
Manufacturer-stated: electrical (ECG) measurement, Bluetooth LE + ANT+, internal memory for one session up to 30 hours, up to 400 h on a CR2025 coin cell, WR30.
[3]Peer-reviewed validation2025
“Impact of Anatomical Placement on the Accuracy of Wearable Heart Rate Monitors During Rest and Various Exercise Intensities”, Sensors (Basel)
n = 28 (14 male, 14 female). Polar H10 chest strap as the reference. Compared Polar Verity Sense (forearm), Garmin Forerunner 55 (wrist) and three simultaneous Whoop 4.0 units at wrist, forearm and upper arm, across rest, cycling warm-up, burpees and a modified Bruce treadmill protocol.
[4]Peer-reviewed validation2024
Robbins et al., “Accuracy of Three Commercial Wearable Devices for Sleep Tracking in Healthy Adults”, Sensors 24(20):6532
Brigham and Women’s Hospital. n = 35 healthy adults, single-night inpatient polysomnography, testing Oura Ring Gen3, Fitbit Sense 2 and Apple Watch Series 8. Authors’ stated limits: one night, lab conditions, healthy adults without insomnia, comparison only during scheduled sleep, and device failures on 6 Apple Watches and 2 Fitbits.